Estimating the Effects of Chronic Medications: Questions to Ask and Tools to Help Answer Them

Background: Traditional analytic approaches are well-suited to the study of point exposures (e.g., a single dose vaccine) but the vast majority of medications are given over a period of weeks, months or years. Decisions about the design and analysis of recurring, multi-dose exposure studies present a significant difficulty for researchers, who will benefit from careful consideration of the interplay of exposures, covariates, and outcomes of interest over time.

Objectives: In this workshop, we will guide attendees through a set of key questions to consider when planning a study of a recurring, multi-dose medication, as well as demonstrate tools that can be used, strategically, to help answer those questions. Researchers wanting additional expertise on the analysis of time-varying exposures and confounding are encouraged to attend.

Description: Using a case study on anticoagulants as a vehicle for the discussion, presenters will lead the audience through each step of the analytic design process. Specific tools to assist with answering critical questions will be presented and discussed: methods decision trees, directed acyclic graphs (DAGs), checklists, and example software modeling implementations in R.